DomTack

Pin any element of a web page for your coding agent: copies its screenshot and full context. No network requests.

DomTack hands your coding agent exactly the element you mean. Click the DomTack button, then click any element of the page: a button, a table row, an error message. DomTack copies its screenshot and a short reference like domtack:k3x9p2ab. Paste it into Claude Code, Codex, Cursor or any other agent. Rich editors get the image; terminals get the reference, which the agent resolves to the element's full context: page URL, CSS selector, HTML, text, position and a snapshot of the page. In browsers driven by Playwright, each pin also records which playwright-cli session it came from, so the agent can keep working in the same browser. Two modes - On its own, the extension copies the screenshot and a text description of the element, ready to paste. - With the free companion command-line tool (npx domtack install, requires Node.js), each pin is saved on your computer and your agent fetches everything with domtack <id>, a Claude Code skill or an MCP server. Private by design - Nothing happens until you click the button; no page is read otherwise. - No network requests, no account, no analytics. Everything stays on your computer. - Open source (MIT): https://github.com/srobfr/domtack Also - History of your pins: right-click the button, then History. Search, copy again, delete. - Keyboard shortcut: Alt+Shift+T. Esc cancels.